Project: Decentralised AI for Situational Awareness 

Key information:

Abstract: 

Robotic swarms are complex systems that are made up of a large number, sometimes hundreds or thousands, of relatively simple robots. They rely on local communication between individual members of the swarm to spread information about their state and the environment in order to make decisions and accomplish tasks such as exploration and mapping. But these communications can be hindered by a variety of factors, for example in a contested environment, adversaries may be jamming communication within an area and the swarm may wish to avoid communication in certain areas to avoid the chance of being detected by those adversaries. When communications break down between members of a swarm, it can become difficult to maintain accurate and up to date information about the state of the swarm and the environment that it is operating in. This problem is brought to the forefront when humans are actively involved in the situation where they may be acting as operators, supervisors or team-mates of the swarm. Here it is vital that the swarm can distil and disseminate the vast amounts of information that it is collecting to the humans in an effective way, in order to provide a human with situational awareness. This project aims to develop techniques that can aid a swarm system’s ability to maintain and improve a human’s situational awareness, even in environments where communications may be limited. These approaches will focus on maintaining information of the current and projected future states of the swarm and the environment and on the communication landscape for the swarm. Swarm systems will use these approaches to make collective decisions about the swarm itself and the environment it is operating in and provide human teammates with reliable and up to date projections of the swarm.